USAID/GUATEMALA PEACEBUILDING 
PROJECT IMPACT EVALUATION
Baseline Report
Prepared under Contract No. GS-10F-0033M/AID-OAA-M-13-00013, Tasking N066
This publication was produced by NORC at the University of Chicago at the request of the United States Agency for International 
Development. It was authored by D. Alex Hughes (University of California, Berkeley), Micah Gell-Redman (University of Georgia), and 
Luis A. Camacho (NORC at the University of Chicago). The views expressed in this publication do not necessarily reflect the views of the 
United States Agency for International Development or the United States Government.
DRG LEARNING, EVALUATION, AND 
RESEARCH ACTIVITY
USAID/GUATEMALA PEACEBUILDING 
PROJECT IMPACT EVALUATION
Baseline Report
July 2020
Prepared under Contract No. GS-10F-0033M/AID-OAA-M-13-00013, Tasking N066
Submitted to:
Brandy Witthoft, USAID COR
Submitted by: 
D. Alex Hughes, Principal Investigator (University of California, Berkeley) 
Micah Gell-Redman, Principal Investigator (University of Georgia) 
Luis A. Camacho, Principal Investigator (NORC at the University of Chicago)
Contractor:
NORC at the University of Chicago
4350 East West Highway, 8th Floor
Bethesda, MD 20814
Attention: Renée Hendley
Tel: 301- 634-9489; E-mail: Hendley-Renee@norc.org
DISCLAIMER 
The authors’ views expressed in this publication do not necessarily reflect the views of the United 
States Agency for International Development or the United States Government.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | i 
TABLE OF CONTENTS
ACRONYMS III
EXECUTIVE SUMMARY 1
INTRODUCTION 6
BACKGROUND 7
IMPORTANCE OF WOMEN IN LOCAL INSTITUTIONS 7
SOCIAL CONFLICT IN THE WESTERN HIGHLANDS 8
RESEARCH QUESTIONS AND HYPOTHESES 9
DIVERSE DECISION MAKERS 9
PROJECT IMPACT 11
INTERVENTION 12
THEORY OF CHANGE 13
IMPACT EVALUATION DESIGN 14
ELIGIBLE COMMUNITIES 14
RANDOMIZATION 14
DATA COLLECTION 16
DO NO HARM CONSIDERATIONS 17
BASELINE DATA COLLECTION 18
FIELDWORK OUTCOMES 19
FIELDWORK CHALLENGES 19
FINDINGS 20
CONFLICT 20
MIGRATION 26
ECONOMIC SITUATION, EDUCATIONAL ATTAINMENT, AND ASSET OWNERSHIP 27
INTERPERSONAL TRUST AND SOCIAL COHESION 30
INTEGRATION ACROSS GENDERS 33
VIEWS ABOUT WOMEN’S LEADERSHIP 36
CIVIC ENGAGEMENT 40
CONCLUSION 42
REFERENCES 44
APPENDIX: POWER CALCULATIONS 46
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | ii
LIST OF TABLES
Table 1. Expected process differences as a result of including women 11
Table 2. Expected outcome differences as a result of including women 12
Table 3. Estimation results for treatment assignment regressions 15
Table 4. Conflict resolution modes/venues 24
Table 5. Conflict impacts 24
Table 6. Migration reasons 27
Table 7. Household asset ownership 29
Table A1. Statistical Power for Individual-level CBPT vs. CBPT+DD Tests 46
Table A2. Statistical Power for Individual-level Any CBPT vs. Control Tests 47
Table A3. Statistical Power for Community-Level CBPT vs. CBPT+DD Tests 48
Table A4. Statistical Power for Community-level Any CBPT vs. Control Tests 49
LIST OF FIGURES
Figure 1. Reported conflict by community 21
Figure 2. Reported conflict by community for four conflict types 22
Figure 3. Proportion of respondents directly impacted by conflict, by conflict type 23
Figure 4. Proportion of conflicts with attempted resolution, by type of conflict 23
Figure 5. Respondents who believe actors are relevant for conflict resolution 25
Figure 6. Respondents who believe actors are effective in resolving conflict 26
Figure 7. Perceptions of change in economic situation 28
Figure 8. Distribution of educational attainment 29
Figure 9. Trust in three reference groups 31
Figure 10. Trust in those living in own community and other communities 32
Figure 11. Responses to Two Hypothetical Crises 33
Figure 12. Social relationships with men 34
Figure 13. Social relationships with women 34
Figure 14. Social interaction with men 35
Figure 15. Social interaction with women 35
Figure 16. Women’s participation in decision making 37
Figure 17. Acceptance of women’s leadership 38
Figure 18. Views about more women’s involvement in leadership 38
Figure 19. Views about women’s concerns being taken seriously 39
Figure 20. Views about women’s leadership ability 40
Figure 21. Inability to attend social and political meetings 41
Figure 22. Reasons for not attending meetings 42
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | iii
ACRONYMS
COCODE Consejo Comunitario de Desarrollo (Community Development Council) 
COMUDE Consejo Municipal de Desarrollo (Municipal Development Council) 
CRA Conflict and Resilience Analysis 
CV Community Vision
DD Diverse Decision Makers Intervention 
DRG Democracy, Human Rights and Governance
GOG Government of Guatemala
IE Impact Evaluation
MDES Minimum Detectable Effect Sizes
RCT Randomized Controlled Trial
TP Tejiendo Paz (local name of Peacebuilding Project)
USAID United States Agency for International Development
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 1 
EXECUTIVE SUMMARY
USAID/Guatemala’s Peacebuilding Project (locally branded as Proyecto Tejiendo Paz, hereafter TP) 
seeks to reduce social conflict and violence, increase social cohesion and trust, and contribute to 
peacebuilding in the Western Highlands Region. The project relies on inclusive, community-led 
engagement, dialogue, and mapping to: (1) identify, prioritize, and develop action plans addressing 
sources of and increasing resilience to social conflict; (2) build partnerships between communities 
and external entities to implement plans prioritized by communities; and (3) strengthen local 
capacity to participate in managing, responding to, and mitigating local conflicts. 
In partnership with USAID’s Center of Excellence on Democracy, Human Rights and Governance 
(DRG), USAID’s Office of Conflict Management and Mitigation, USAID/Guatemala, and the project 
implementer, Creative Associates International, NORC at the University of Chicago is conducting an 
impact evaluation (IE) to assess the overall effectiveness of project activities at the community level. 
The IE tests the key hypothesis that underpins the project’s theory of change, namely that 
community-based initiatives to address conflict drivers can reduce social conflict and increase social 
cohesion. In addition, the IE assesses whether promoting increased women’s participation in 
community deliberations can further contribute to reducing conflict and increasing cohesion.
To achieve these objectives, NORC’s IE team randomly assigns 195 communities eligible to receive 
programming into one of three arms in equal numbers (65 communities each): 
• TP, where communities receive the core project’s programming as designed by 
USAID/Guatemala and the implementer; 
• TP+DD, where communities receive the core project’s programming plus an additional diverse 
decision makers intervention promoting increased women’s participation; and
• Control, where communities do not receive any programming. 
Randomization allows us to measure the impact of TP as designed, as well as the marginal effects of 
the diverse decision makers intervention. Diverse decision makers focuses on reducing the barriers 
to participation for women through a training program called women peace promoters (asesoras de 
cambio). The asesoras de cambio intervention is based on the premise that to create sustainable 
gender equality and lasting peace, women need to have increased capacity for civic participation and 
political mobilization.
As designed by TP’s implementer, the asesoras de cambio will be identified through a rapid 
stakeholder assessment to be undertaken in all TP communities. Selected women will include 
community leaders, COCODE members, mothers and grandmothers, youth, midwives (comadronas),
and ancestral authorities (Ajq’ij). TP will build their capacity to lead so that they can take on more 
active roles in community decision making and in supporting the resolution of social conflict. The 
asesoras de cambio will receive training on topics such as leadership, decision making, effective 
communication, conflict transformation, negotiation, critical reflection and dialogue, political 
expression and participation, and women’s legal rights.
In the short term, we expect the diverse decision makers intervention to change decision-making 
processes and outcomes above and beyond core TP programming. We also expect to observe some 
changes in individuals’ attitudes and beliefs, especially those related to the role of women in decision￾making processes and the legitimacy of TP deliberation fora. In the long term, we expect the diverse 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 2 
decisions makers intervention will contribute to a decrease in social conflict and an increase in social 
cohesion above and beyond core TP programming.
Data collection consists of a three-wave (baseline, midline, and endline) panel survey of individuals 
16 years old and over. Our survey aims to capture the opinions and perspectives of youth—one of 
the key focus population of TP. Recognizing the challenges associated with obtaining informed 
consent from minors, we decided to limit the underage population included in the survey to 16-year￾olds and 17-year-olds. The primary purpose of these surveys is to measure the incidence of conflict 
and the level of cohesion within communities, in a way that is identical between treatment and 
control. The questionnaire includes modules related to conflict, social cohesion and trust, integration 
across genders, responses to disasters, development and poverty, migration, views about community 
and regional leadership, women’s leadership, youth’s leadership, other political attitudes and 
behaviors, and personal and household socioeconomic characteristics.
Baseline fieldwork took place between September 10 and October 27, 2019. According to the 
evaluation design, midline measurement will take place immediately after the completion of 
community vision development activities in each community, and endline measurement will occur 
near the end of the project in 2023.
At baseline, teams attempted to conduct interviews in all 195 communities, but they did not receive 
authorization to carry out work in 11 communities in the Quiché Department. In most cases 
community authorities were unwilling to participate because participation did not entail any material 
benefits to the community. Another reason authorities gave was that their participation in past 
studies had not yielded any results for them. In addition, the authorities of two communities—
Pocohil Primero and Mactzul Sexto in the Chichicastenango Municipality—asked enumerators to 
immediately leave their communities, threatening them with violence if they failed to do so. The data 
collection effort resulted in a dataset containing 3,866 valid cases. The most challenging aspect of 
fieldwork activities was scheduling, as several authorities changed the dates in which they would 
allow visits to their communities with little advance notice.
Below we summarize the most important baseline findings:
CONFLICT
• To measure conflict incidence, we ask respondents whether they are aware of conflict having 
occurred within their community in the last 12 months for each of four separate conflict types: 
land; water or water access; natural resources; and governance, politics, or municipal or 
community authorities.
• Respondents report at least one conflict having occurred in the majority of communities. 
However, conflict does not affect all people within communities equally. In the majority of 
communities, fewer than five respondents report conflicts; in many communities, between five 
and ten respondents report conflicts; and in a smaller but substantial number of communities, 
more than ten respondents report conflicts. 
• We find similar patterns when we analyze the four specific types of conflict we ask about—land, 
governance, natural resources, and water—separately. That is, in almost all communities, at least 
some respondents report conflict, but only in a minority of communities does a large share of 
respondents report conflict.. The most common types of conflict are those over land and water.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 3 
• The majority of respondents who report a conflict say that they are not aware of the parties in 
conflict having attempted resolution. In the minority of instances in which resolution attempts 
are reported, by far, the most common modes/venues for attempted resolution are formal 
institutions at the community and municipal level, along with the courts.
• To gauge the extent to which conflict impacts daily life, we ask respondents if conflict had forced 
them to avoid certain behaviors or activities (e.g., travelling outside their community and 
attending community events) or exposed them to some negative experience (e.g., being 
threatened with physical violence and suffering property damages or losses). Due to space 
constraints we only asked general questions about all conflict, rather than about each of the four 
specific types of conflict mentioned above. 
• The most common specific conflict impacts on our respondents’ behaviors and experiences are 
children not attending school and respondents avoiding leaving their homes alone at night.
MIGRATION
• To gauge migration past experiences and prospects we ask a number of questions, including 
place of origin, plans to migrate abroad and within Guatemala, including desired destination, and 
reasons for wanting to migrate.
• Roughly one third of our respondents have not lived in the community in which they were 
interviewed for their entire lives. Of these, more than two thirds come from another part of 
Guatemala, and less than one third come from abroad.
• Roughly 20 percent say that in the past 12 months they have considered moving to another 
country. Of these, nearly all say their intended destination is Mexico. Fifteen percent say they 
considered moving to another part of Guatemala. Interestingly, the proportions of those 
reporting an intention to migrate corresponds closely to the proportions of those reporting 
having a household member who has migrated in the past 12 months.
• Economic motivations emerge as the most widely reported motivation for plans to migrate.
INTERPERSONAL TRUST AND SOCIAL COHESION
• To gauge interpersonal trust, we ask our respondents to state how trustworthy (1) their 
neighbors; (2) those who live in their community; and (3) those who live in other communities 
are. The most trustworthy group is those who live in the respondents’ communities. Those who 
live in other communities are the least trustworthy group and respondents’ neighbors are 
somewhere in between.
• Individuals’ feelings of trust are complex. Respondents can be trusting of people living in their 
own communities while not being trusting of those living in other communities. We see that 
respondents who are very trusting of those living in other communities also tend to be very 
trusting of their own community. Respondents who do not trust those living in their 
communities are not likely to trust those living in other communities.
• To gauge social cohesion, we ask our respondents to imagine two hypothetical crises, a negative 
act of god (a flood, landslide, or earthquake) and a threat that affects mostly women in the 
community (the community is left without a midwife), and answer three questions about how 
the community would respond. The questions ask whether the response would be along familial 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 4 
lines or encompassing of the community as a whole; along gender lines—separate male or 
female responses, or all-inclusive responses; and along age lines—separate youth or older adult 
responses, or all-inclusive responses.
• We find that the bulk of respondents say that their communities would respond by working 
together as a whole—rather than segregated along familial, gender, or age lines. 
INTEGRATION ACROSS GENDERS 
• Motivated by patterns of discrimination against women and gendered segregation in the region, 
we ask questions to gauge the integration of men and women in these communities.
• The preponderance of respondents say that they get along either “well” or “very well” with the 
men and women in their communities. However, there is evidence for gender-based responses; 
male respondents are more likely than female respondents to get along “very well” with other 
men, while female respondents are more likely than male respondents to get along “very well” 
with women.
• When it comes to interaction with people of a specific gender, the most frequent response is 
that individuals spend social time with others once per week. Patterns also show that individuals 
spend more time with others who share the same gender identity: men more frequently report 
spending social time with other men daily and weekly than do women, and women more 
frequently report spending social time with other women daily and weekly than do men.
VIEWS ABOUT WOMEN’S LEADERSHIP
• Views about women’s leadership are quite positive. The vast majority of respondents agree with 
the statements “Women in my community participate in making decisions that affect the whole 
community;” “Women are accepted in leadership roles in your community;” “There should be 
more women’s leadership in your community;” “When women in my community raise concerns, 
they are taken seriously;” and “Women in my community are as capable of being leaders as 
men.”
• There are some differences in views about women’s leadership across gender, with the views of 
females being slightly more positive than those of males. Overall, however, differences are not as 
stark as one would expect in a context where women have traditionally been excluded from 
leadership positions.
• We also ask respondents to share their views on the barriers to women’s participation in 
community governance. The reasons that most respondents strongly agree with are lack of 
safety and respect for women in the political sphere; formal laws and policies preventing the 
participation of women; lack of education and literacy among women; and women’s greater 
responsibilities at home. Difference across genders in these responses are modest.
CIVIC ENGAGEMENT
• Our survey also includes a number of questions gauging participation in meetings of various 
types of social and political groupings or organizations. We follow these questions by asking 
respondents whether they would have wanted to attend meetings but were unable to.
Approximately one-third of those surveyed answered affirmatively.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 5 
• We ask individuals who said they were unable to attend to provide a reason. The most common 
reason is lack of time. Remarkably, more men than women report that work at home kept them 
from participating. More women than men report that it would be inappropriate for them to 
participate and that it might be dangerous for them to participate.
As of the writing of this report, the IE has been cancelled. We understand that current funding 
restrictions prevent USAID/Guatemala from contributing funding to IE activities in the short term.
Our hope is that additional funding for the IE becomes available toward the end of the project in 
2023, in time to carry out endline data collection, analysis, and reporting activities. Should the 
implementer abide by the full randomization schedule and additional funding become available, we 
would be able to estimate the impact of the asesoras de cambio intervention by comparing 
communities assigned to TP+DD with communities assigned to TP. We would also be able to 
estimate the impact of the standard TP project by comparing communities assigned to TP with 
communities assigned to the control group.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 6 
INTRODUCTION
USAID/Guatemala’s TP seeks to reduce social conflict and violence, increase social cohesion and 
trust, and contribute to peacebuilding in the Western Highlands Region. The Western Region is 
currently experiencing a blight of social conflict generated by historical patterns of structural 
exclusion, internal armed conflict, and a lack of social cohesion. The 2015 comprehensive assessment 
of conflict in the region revealed patterns of intra-family violence, social conflict between the young 
and old, and conflict among the youth who lack livelihood opportunities. If left unaddressed, social 
conflict is known to decrease development; weaken local, social, and formal governance groups; and 
could ultimately diminish democratic governance.
To address these issues, TP relies on inclusive, community-led engagement, dialogue, and mapping to 
identify, prioritize, and develop action plans addressing sources of and increasing resilience to social 
conflict; build partnerships between communities and external entities to implement plans prioritized 
by communities; and strengthen Government of Guatemala (GOG) and non-governmental capacity 
to participate in managing, responding to, and resolving local conflicts.
TP’s theory of change is:
If the necessary conditions for constructive and inclusive community engagement are 
generated and sustained, if diverse sectors and stakeholders within Western Highlands 
communities can identify and work together to address social conflict at the familial, 
community, and municipal levels, if statutory and traditional authorities have increased capacity 
and understanding of local dynamics to efficiently and effectively respond to local conflicts in a 
transparent and fair manner, as well as improved communication and coordination
mechanisms, then relationships will be strengthened and trust increased within communities, 
between communities, and between communities and other stakeholders (including traditional 
and statutory institutions), communities will increase cooperation and be able/willing to 
resolve conflicts peacefully, and communities will build reliance on existing, recognized, and 
legitimate mechanisms to reduce social conflict, thereby improving social cohesion and 
peacebuilding, and reducing social conflict and violence.
In partnership with USAID’s Center of Excellence on DRG, USAID’s Office of Conflict Management 
and Mitigation, USAID/Guatemala, and the projects implementer, Creative Associates International, 
NORC is conducting an IE to assess the overall effectiveness of TP activities at the community level. 
The IE tests the key hypothesis that underpins the project’s theory of change, namely that 
community-based initiatives to address conflict drivers can reduce social conflict and increase social 
cohesion. In addition, the IE assesses whether an intentional focus on the participation of women—a 
group who has been historically, politically, and socially excluded in the Western Highlands and 
beyond—in community deliberations can further contribute to reducing conflict and increasing 
cohesion. Learning that mandating women’s inclusiveness in peacebuilding activities increases the 
participation in, output of, and ultimately the likelihood of success of those activities would both 
validate previous scientific findings in a new context and provide compelling evidence that relatively 
straightforward program changes may improve program performance across the portfolio of USAID 
projects.
To achieve these objectives, NORC’s IE team randomly assigned 195 communities into one of three 
arms: (1) communities receive the core TP programming (TP); (2) communities receive the core TP 
programming and the additional diverse decision makers intervention (TP+DD); and (3) communities 
receive no programming (i.e., these are pure control communities). 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 7 
The remainder of this report is organized as follows. In the first section we (1) review evidence 
about the importance of local formal and informal institutions for development and peacebuilding, 
and the consequences of including women in decision-making processes. In the first section we also 
provide a brief overview of social conflict in the Western Highlands. In the third section, we present 
the study’s research questions and hypotheses pertaining to both the DD intervention and the 
overall TP project. In the fourth section, we describe the DD intervention named asesoras de cambio
and its associated theory of change. We present the evaluation framework to estimate the 
effectiveness of TP and TP+DD in the fifth section. The sixth section provides an overview of 
baseline data collection activities, including fieldwork outcomes and challenges. We describe our 
study sample along various characteristics, including key outcomes, in section seven. We conclude by 
laying out a plan for positioning the project to conduct IE follow ups should additional funding 
become available.
BACKGROUND
IMPORTANCE OF WOMEN IN LOCAL INSTITUTIONS
Obstacles to collaboration within communities are varied, but can include corruption, the presence 
of ethnic diversity that is not reflected in governance organizations, and economic and gender 
inequality. Improving the ability of communities to pursue common goals by directly addressing these 
hindrances may unlock the potential for peacebuilding that would otherwise be wasted. One 
promising method to pursue these goals is to change the institutional arrangements that govern 
community life.
Institutions—whether formal government-sanctioned bodies or informally constructed—are 
powerful tools for shaping human behavior. When institutions are poorly functioning, unfair, or 
opaque they can act as barriers to peace and prosperity. In contrast, when institutions function 
efficiently, justly, and with the oversight and consent of those individuals whom they represent, they 
can serve as a boon to local peacebuilding.
While there is no generally accepted panacea to improving institutional performance, the scholarship 
largely finds that greater transparency and community involvement improves the quality and function 
of institutions (Ostrom, 1990). Through transparency, participants and the institution can clearly 
define expectations for behaviors of both the institution and those influenced by it; through 
community involvement, collective-choice agreements come to represent the aims of the whole 
community. Even more, when the whole of the community is engaged within a transparent, engaging 
institution, it is likely that members from diverse parts of the community—parts that may be 
involved in forms of social conflict—will work closely to pursue a common goal. Across a diverse set 
of locations and community identities, experimental research has shown that working toward a 
common goal generates mutual feelings of well-being among members engaged in work, and thereby 
reduces feelings of social conflict between diverse community sectors (Allport, 1954; Bornman & 
Mynhardt, 1991; Eskilson 1995; Pettigrew 1998; Broockman & Kalla 2016). 
Chattopadhyay and Duflo (2004) find that including women in village councils in India produced 
sweeping positive effects. When women are included in the decision-making process, the types of 
public projects undertaken by local institutions measurably shift. Women who live in the region 
Chattopadhyay and Duflo studied are primarily responsible for providing water to households, a task 
that in many areas consumed several hours of women’s’ days. Communities that were randomly 
assigned to include women in the decision-making process made measurably larger investments in 
public goods that secured water access—a direct result of the randomized controlled trial (RCT) 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 8 
that included female decision makers. Even more strikingly, ten years after the RCT began, effects of 
including female decision makers continued to be measurable: areas that had been assigned female 
council leadership were more likely to elect female leaders (Chattopadhyay and Duflo, 2009). What 
is more, female leadership also increased career aspirations among women (Beaman, Duflo, Pande, 
Topalova, 2012) through role-model effects.
Olken (2010) finds two key results for how public decision-making causes changes in outcomes. 
First, participatory decision-making bodies caused higher satisfaction with outcomes and process (in 
contrast to more common elected-representative bodies). Second, when women were involved in 
the decision-making process, the development activities hued toward women’s priorities, specifically
toward expenditures on the development of water provision resources.
We draw two primary conclusions from the literature. First, participatory bodies (like those 
required by all forms of TP) cause increased satisfaction in the development process. Second, the 
inclusion of women in the decision-making process causes the policy to reflect preferences of 
women. Importantly, most experimental research has focused on development projects rather than 
on peacebuilding projects. This IE therefore seeks to contribute to improving our understanding of 
the effects of including women in peacebuilding processes. The 2018 United Nations Secretary￾General Report on Women and Peace and Security identifies several of the potential benefits of 
including women as well as the relatively weak, and potentially decreasing role that women and 
other non-male genders play in peace agreements.
Below we lay out a strategy to evaluate whether TP, through its equality and local institutions focus, 
can both increase the performance of local institutions and reduce social conflict. Ultimately, success 
for TP would consist of demonstrable changes in the attitudes and behaviors that lead to social 
conflict, a change in the beliefs about how to resolve conflict, and eventually decreased occurrence 
of conflict events. We also lay out a strategy to assess whether deliberate inclusiveness of women in 
peacebuilding activities increases the effectiveness of TP.
SOCIAL CONFLICT IN THE WESTERN HIGHLANDS
As reported in the Guatemala Conflict Vulnerability Assessment (Adams, 2015), the Western 
Highlands region has witnessed an ongoing legacy of systematic social conflict. Conflict, sometimes 
rising to the level of physical violence is endemic to nearly all aspects of society in the Western 
Highlands region. Conflicts include intra-family violence including violence against women and 
children, clashes between older and younger generations within families, and political rifts that 
quickly transcend the family and manifest at the community level; governance conflicts related to 
gangs1 and to para-state governance mechanisms emerging to cope with this and other problems; 
and sometimes long-standing conflicts related to land and other natural resources, climate change, 
and extractive industries.
 
1 The Guatemala Conflict Vulnerability Assessment is deliberate in its description of youth-groups as “gangs,” 
noting that “in most communities [they] are seldom visible… and are not killing people.” These groups of 
organized youths “have emerged […] as a result of young people’s response to internal community and family 
dynamics” (Adams, 2015, p. 14). These groups of youths are unaffiliated with notorious urban Guatemalan and 
Salvadoran gangs (e.g. Mara Salvatrucha and Mara 18) and instead the term is used locally as a catch-all term. 
The Conflict Vulnerability Assessment notes that, “Gangs in many rural communities (at the time of [the] 
study) were very different from the more virulent urban gangs in Guatemala City, but they all were cast 
erroneously as the same dangerous thing” (Adams, 2015, p. 14).
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 9 
The Conflict Vulnerability Assessment finds that the conflict, if left unresolved, stands the possibility 
of further escalation to the point that it may realize active, violent conflict. However, even without 
the possible outset of violent conflict, the unresolved, simmering social conflict has deleterious 
effects on community structure and peacebuilding. What is more, recent research finds that 
longstanding conflict harms individual physical, emotional, and mental health. Thus, unaddressed 
social conflict poses a serious threat, possibly escalating into physical violence, but also causing 
worse human outcomes and threatening democratic stability.
Although the Western Highlands Region is experiencing broad-based, systematic social conflict, 
perhaps the leading finding from the Conflict Vulnerability Report is that the specific modalities of 
this conflict vary widely from community to community, and in some cases, even from family to 
family. While acknowledging that multiple sources of conflict create challenges from the perspective 
of the intervention and evaluation design, it also potentially allows peacebuilding interventions to 
specifically target the causes and forms of conflict felt most acutely in each community. As an 
example, several communities in Quiché have identified that the primary source of conflict is 
significant low-level criminal “gang” activity, whereas in San Marcos the leading cause of conflict 
concerns the implementation of hydropower. It seems quite evident that programming that seeks to 
ameliorate one form of these two conflicts would be poorly posed to address the other form.
Despite the varied experience of conflict, the Conflict Vulnerability Report suggests that there are 
systematic groups or types of conflict modalities. These groups, which the report identifies as 
generally including (a) intergenerational conflict, (b) employment conflict, (c) land-rights conflict, and 
(d) extractive enterprises conflict, inter alia, mean that TP and the IE team can plan to work with 
communities experiencing a relatively constrained set of conflict considerations.
RESEARCH QUESTIONS AND HYPOTHESES
TP seeks to strengthen communities by including all community members in efforts to identify, 
prioritize, address, and resolve sources of social conflict and violence. By addressing these issues, TP 
aims to increase social cohesion, reduce conflict, and promote peace. Calls for inclusive 
programming are present across a broad range not only of the USAID portfolio but also of a diverse 
set of peacebuilding and development organizations. While there is considerable evidence for the 
importance of inclusion in the case of development, the evidence is underdeveloped in the case of 
peacebuilding interventions. This IE is designed to create a specific contrast in women’s ability to 
participate in conflict identification through the DD intervention to be implemented in a subset of 
TP communities. The IE will examine which parts of the conflict-reduction and peacebuilding process 
respond to a focus on inclusivity of women.
DIVERSE DECISION MAKERS 
Research Question 1: What is the effect of including diverse sectors and stakeholders in the 
process of community-led peacebuilding?
There are four pathways by which increasing the diversity of local decision makers might reduce 
conflict or increase cohesion. We report these in Table 1.
First, including previously excluded women’s groups could change the preferences of local 
leaders. The core idea here is that members of excluded groups might simply pursue different goals 
than their counterparts. Perhaps the strongest evidence for this comes from India, where mandating 
women’s participation in local governing councils improved the quality of drinking water 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 10
(Chattopadhyay and Duflo, 2004) and made it more likely that women would achieve political 
representation, even after the quotas were removed (Bhavnani, 2009). While studies using 
observational data find less clear-cut evidence (Franceschet and Piscopo, 2008), there seems to be 
consensus that, at least on some issues, women do govern differently (Piscopo, 2014).
Second, including previously excluded groups could change the beliefs about the role of the 
community population in the practice of peacebuilding and decision making. Contact theory (Allport, 
1954) holds that when groups of people engage in a common purpose, feelings of mutual well-being 
increase and feelings of negative sentiments towards others decrease. The benefits of this contact 
hypothesis have been demonstrated in broad and varied circumstances. Recently, Broockman and 
Kalla (2016) showed that even incidental contact between out-groups and dominant groups in the 
United States have the power to lastingly improve attitudes and beliefs toward those out-groups. 
Pettigrew (1998) identifies a litany of examples of successful blunting of social conflict as a result of 
intergroup contact: interracial works in South Africa (Bornman and Mynhardt, 1991); white school 
children coming to know Southeast Asian immigrants (Wagner et al., 1989; McKay and Pitman, 1993; 
Riordan, 1987); and heterosexual people coming to know homosexual people (Eskilson, 1995; Herek 
and Capitanio, 1996). Improvements in intergroup attitudes and beliefs have been demonstrated 
across a remarkably broad set of circumstances.
Third, increasing the representation of excluded groups might change who participates in 
decision-making processes. Evidence from India on this point is particularly compelling. Increasing 
greater ethnic/economic diversity in India led to a greater level of transfers to groups that had 
previously been under-represented (Pande, 2003), possibly by increasing the influence of those 
voters. Also in India, increasing women’s representation led to an overall increase in women’s 
participation in politics (Bhavani, 2009). Note that, while increasing the participation of a previously 
excluded group may seem unambiguously beneficial, it brings with it a potential unintended 
consequence. A more diverse set of preferences can impede coordination, making it difficult for 
communities to provide basic goods that benefit all residents (Alesina, Baqir, and Easterly, 1999). 
Our intervention and outcome measures should be designed with this theoretical ambiguity in mind.
Fourth, increasing diversity might change the dynamics between members of the legislative or 
consultative body itself. For example, when united with a common purpose and closely working 
together, members of a consultative body may develop new and positive opinions toward members 
with whom they previously conflicted. Changing the dynamics within the legislative or consultative 
body is perhaps the most controversial of the four pathways, and the one for which evidence from 
the field is the most ambiguous. The literature in social psychology and organizational behavior has 
indicated a number of possible mechanisms linking diversity to leadership outcomes, but the 
evidence for many of these is inconsistent. Moreover, very few of these theories have been tested in 
field settings in developing countries. Perhaps the most pertinent example is a study of community￾driven development in Liberia, (Fearon et al., 2015), which found that increasing community 
participation in monitoring development led to greater cooperation only among leadership groups 
with gender diversity.
While the theory linking this fourth mechanism to conflict and cohesion is the least developed, it has 
significant potential to improve peacebuilding policy. Diversity quotas are widespread, but 
understanding their impact at the local level remains limited. Our intervention can make a significant 
contribution to this area of knowledge by collecting outcome measures that provide a detailed 
picture of how internal leadership dynamics change in response to increased diversity.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 11
Table 1. Expected process differences as a result of including women
HYPOTHESIS MEASUREMENT 
Changing Preferences • Distinct conflict modes identified by community
• Differences in best response to conflict
Changing Beliefs • Changes in beliefs about appropriateness of women’s participation
• Change in beliefs about nature of underlying conflict 
Changing Participation • Turnout of women at community meeting 
• Participation of women in community vision creation
• Participation of women in execution of community vision
Changing Dynamics • Increased active participation by women 
• Change nature of conversation among men • Increased women’s participation beyond community-led peacebuilding
PROJECT IMPACT
If inclusive programming changes the process of conflict resolution and peacebuilding, we anticipate 
that it will also have measurable impacts in terms of overall project impact. Each change in 
preferences, beliefs, participation, and dynamics identified in the previous subsection points to 
increases in legitimacy and the development of a broad base of support for community deliberation 
fora. These community deliberation fora—whose successful functioning is a primary outcome of 
Objective 1 in the project description—aim to build resilience by increasing communities’ capacity 
for managing and resolving conflict.
Research Question 2: What is the effect of including diverse sectors and stakeholders in the 
outcomes of community-led peacebuilding?
When diverse actors work collaboratively and in good faith to solve a common problem, the work 
serves to humanize the out-group, reducing negative feelings between groups, increasing social 
cohesion, and reducing social conflict (Allport, 1958, p. 264). Even more, recruiting diverse groups of 
individuals to be a part of decision-making bodies has been shown to increase the provision of goods 
to minority groups (Chattopadhyay and Duflo, 2004). In the context of an inclusive community 
peacebuilding project, the provision of goods to whole communities rather than more limited 
sectors within a community is of paramount importance.
We identify three social cohesion and peace outcomes that may change by targeting diverse decision
makers. 
First, citizens may view projects as being increasingly legitimate because the conflict assessment 
and initial conflict identification involved a broader set of participants. Normative accounts of 
political legitimacy distinguish between a de facto political body and one whose existence is 
somehow justified (Rawls, 1993). Broadening the basis of who identifies conflict is likely to increase 
the perceived justification for the existence of TP’s community visions (CVs) and subsequent conflict 
resolution outcomes.
Second, communities that start project activities with a diverse set of decision makers may be 
increasingly capable of undertaking action and implementing CVs. In communities where the 
process is more legitimate, community members will be more willing to undertake the costs and 
actions necessary to make CVs thrive. This increased contribution and resulting capacity might be 
measurable in terms of conflicts resolved, number of individuals involved in CV implementation, or 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 12
successful requests for resources made to municipal-level authorities, like Consejos Municipales de 
Desarrollo (COMUDE).
Third, project activities undertaken by communities with a diverse set of leaders may result in 
increased peace and social cohesion. More legitimate activities are likely to be more effective in 
resolving issues that arise from social conflict and fostering social cohesion. At the most acute levels, 
we might expect that communities where women are involved in the decision-making process would 
experience lower rates of physical violence in general, and especially against women in those 
communities. As the 2015 conflict assessment makes clear, though, many forms of conflict in these 
communities are not rising to the levels of physical conflict. In these areas, contingent on the type of 
community conflict, we might also expect improvements in measured levels of non-violent conflict.
We summarize these expectations in the table below:
Table 2. Expected outcome differences as a result of including women
HYPOTHESIS MEASUREMENT 
Increased 
Legitimacy
• More positive attitudes and beliefs about the legitimacy of community decision￾making fora
• More positive attitudes about community orientation toward social conflict
Increased 
Capacity • Number of conflicts resolved through the community vision
• Results requests made and/or granted to Municipal authorities
Increased 
Peace and Social 
Cohesion
• Reduced feeling of dis-confidence broadly through community
• Reduced feeling of dis-confidence especially in diverse group • Reduced recorded events of violence 
• Reduced presence of out-migration
INTERVENTION
TP aims to create community-led action to reduce social conflict and create resilient social and 
formal institutions to address the sources of conflict. Furthermore, TP expressly aims to engage 
historically marginalized groups in that process. And so, it is natural for this IE to assess the 
effectiveness of deliberately broadening the set of people who are involved in community-led 
intervention activities.
In their present state, decision-making bodies in communities (frequently the COCODES) include 
very little gender diversity—nearly all seats are held by male members of the community. In many 
communities, women hold positions in semi-formal auxiliary committees, but these committees lack 
the authority and funding that is attendant to the formal groups.
The core TP approach focuses on involving diverse sets of stakeholders in ad hoc community 
stakeholder meetings where they participate in identifying conflicts and developing CVs to address 
conflicts. The diverse decision makers version of TP (TP+DD) focuses on reducing the barriers to 
participation for one particular marginalized group—women. To support women’s inclusion in 
stakeholder meetings, TP will undertake an intervention called women peace promoters (asesoras de 
cambio) in 65 TP+DD communities. The asesoras de cambio intervention is based on the premise that 
to create sustainable gender equality and lasting peace, women need to be given the tools to 
increase their potential for civic participation and political mobilization. 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 13
As designed by TP’s implementer, the asesoras de cambio will be identified through a rapid 
stakeholder assessment to be undertaken in all TP communities. Selected women will include 
community leaders, COCODE members, mothers and grandmothers, youth, comadronas, and 
ancestral authorities (Ajq’ij). TP will build their capacity to lead so that they can take on more active 
roles in community decision making and in supporting the resolution of social conflict. The asesoras 
de cambio will receive training on topics such as leadership, decision making, effective 
communication, conflict transformation, negotiation, critical reflection and dialogue, political 
expression and participation, and women’s legal rights.
The main responsibilities of the asesoras de cambio as initially envisioned by the project team include:
• Create women’s groups in their communities, with a focus on supporting the implementation of 
the Conflict and Resilience Analysis (CRA) and contributing to the development and 
implementation of CVs.2
• Serve as a liaison between the project and other community stakeholders, particularly other 
women. 
• Promote the active participation of women in peacebuilding and civic processes.
• Replicate training received with other groups of women.
THEORY OF CHANGE
The key component of the theory of change we identify in this IE is derived from 
USAID/Guatemala’s theory of change:
If diverse sectors and stakeholders within Western Highlands communities can identify and 
work together to address social conflict at the familial, community, and municipal levels, then
relationships will be strengthened and trust increased within communities, thereby improving 
social cohesion and peacebuilding, and reducing social conflict and violence.
Dialogues that recognize conflict necessarily reflect the beliefs and understanding of dialogue 
participants. If a dialogue were to include only a single, homogeneous group of actors, the identified 
conflict and its remedy would be highly likely to reflect the beliefs of that single group of actors.
TP program activities expressly focus on bringing diverse sets of stakeholders to participate in these 
dialogues. However, community/contextual challenges might mean that some stakeholders, despite 
interest in participating, may be unable to make their voice heard. Our intent in the IE, then, is to 
provide capacity-building training and explicitly invite women leaders to engage in forming the 
conflict-understanding dialogue and peacebuilding program (i.e., in the CRA and CV development 
processes) through the asesoras de cambio intervention. Importantly, we do not view this as a 
distinct theory of change from that of the core program. Instead, we view this IE as testing one of 
 
2 According to project documentation, the CRA analysis will identify emerging sources and drivers of conflict, 
potential triggers, and mitigating factors; primary actors and their roles in driving and mitigating conflict; 
sources of resilience; existing conflict management processes; communication patterns; anticipated challenges 
and opportunities; and external influences. The CRA thus serves as the main source of information for the 
subsequent development of CVs.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 14
the core hypotheses of the theory: what happens if program activities remove barriers to 
participation for a historically marginalized group?
IMPACT EVALUATION DESIGN
The methodology for this IE involves three steps. First, TP’s implementer identified a set of 
communities to be considered for programming and evaluation. Second, the IE team randomly 
assigned 65 communities to each of the following groups: TP (standard project); TP+DD (standard 
project plus asesoras de cambio intervention); and control (no programming). Third, the IE team will 
collects data at three points to estimate the impact of the intervention: (1) at baseline, before the 
start of project activities; (2) at midline, immediately after the completion of community vision 
development activities in each community; and (3) at endline, near the end of the project in 2023. 
We describe these three steps below followed by “do no harm” considerations. 
ELIGIBLE COMMUNITIES
TP’s implementer identified 195 communities eligible for programming in 15 municipalities in the 
Western Highland. Five districts are located in Huehuetenango, six districts in Quiché; three 
districts in San Marcos; and one district in Totonicapán. In line with TP’s Cooperative Agreement, 
the implementer selected the departments and municipalities in coordination with 
USAID/Guatemala. Next, the implementer conducted meetings with representatives from the 356 
communities located in geographic areas where the project could operate and collected data to 
score these communities along three selection criteria: (1) presence of conflict and vulnerability, (2) 
social and political will to work with the project, and (3) potential and opportunities for project 
impact. Finally, the implementer combined this information using an eligibility index, ranked 
communities according to the index, and made the final selection of 195 communities.
RANDOMIZATION
Communities were randomly assigned to the three groups—TP, TP+DD, and control—using block 
random assignment. The IE team used information from TP’s community selection database to 
create groupings (referred to as blocks in the technical language of the IE) of similar communities 
within each municipality. Within these blocks of communities, the IE team assigned one community 
to each of the intervention conditions. Blocked randomization, as it was conducted in this 
intervention, ensures the randomization produces groups of 65 communities that are highly similar 
on measurable characteristics except for the differences in the intervention they receive. Blocked 
randomization maximizes statistical power and the IE’s ability to measure differences in outcomes 
due to program activities.
The IE team used information from TP’s community selection database that is predictive of conflict 
and program uptake to conduct the block randomization:
• Women’s Participation in Governance – How frequently do women participate in governance?
• Total Problem Frequency – How frequently are there problems in the community?
• Problems Due to Natural Resources – How frequently do problems related to natural resources 
(e.g., hydroelectric dams or forestry) occur in the community?
• Problems Due to Land Use – How frequently do problems related to land-use occur in the 
community? 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 15
• Existence of Reported Domestic Violence – How frequently does domestic violence occur in the 
community? 
• Community Plans for Conflict Resolution – Does the community have a plan to respond to problems 
that might arise?
• Community Organized for Conflict Resolution – Is there a community organization to resolve 
problems? 
• Total COMUDE/Community Representative Support – How many of the working groups of 
COMUDE/community representatives at community selection meetings recommended that this 
community be selected for inclusion in the program? 
We checked for covariate balance (i.e., checked whether the randomization was successful in 
creating balance between the communities assigned to different treatment conditions) using the 
current standard procedure: a test to predict treatment assignment using available covariates 
(Gerber and Green, 2012). In particular, we estimated three models of the following form: 
where there is one model estimated for each of the three intervention arms (control, TP, TP+DD). 
One can conclude randomization was successful when it is not possible to predict treatment with 
data on hand. As we report in Table 3, that was the case here. The probabilities for all tests are 
higher than 0.05, so we conclude that there is no evidence to suggest an unsuccessful randomization.
Table 3. Estimation results for treatment assignment regressions
DEPENDENT VARIABLE:
ASSIGNED CONTROL
(1)
ASSIGNED TP
(2)
ASSIGNED TP+
(3)
Women Participate In 
Governance
0.020
(0.052)
0.038
(0.052)
-0.058
(0.052)
Community is Organized 0.004
(0.005)
-0.007
(0.005)
0.004
(0.005)
Community has a Plan 0.026
(0.039)
-0.012
(0.038)
-0.014
(0.038)
Total Mesa 
Recommendations
0.014
(0.038)
-0.035
(0.038)
0.021
(0.038)
Total Problem Frequency -0.008
(0.042)
0.027
(0.042)
-0.020
(0.042)
Problems due to Natural 
Resources
-0.024
(0.043)
-0.010
(0.042)
0.035
(0.043)
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 16
DEPENDENT VARIABLE:
ASSIGNED CONTROL
(1)
ASSIGNED TP
(2)
ASSIGNED TP+
(3)
Problems due to Land 0.018
(0.050)
-0.0001
(0.050)
-0.018
(0.050)
Constant 0.265
(0.186)
0.346*
(0.185)
0.389**
(0.185)
Observations 195 195 195
R2 0.009 0.020 0.015
Adjusted R2 -0.028 -0.017 -0.022
Residual Std. Error (df = 187) 0.479 0.477 0.478
F Statistic (df = 7; 187) 0.238 0.543 0.414
Note: *p<0.1; **p<0.05; ***p<0.01
DATA COLLECTION
Data collection consists of a three-wave (baseline, midline, and endline) panel survey. The primary 
purpose of these surveys is to measure the incidence of conflict and the level of cohesion within 
communities, in a way that is identical between treatment and control. Baseline data collection took 
place between September and October 2019. Midline measurement will take place immediately after 
the completion of community vision development activities in each community. This measurement is 
particularly important to gauge the expected impacts of the DD intervention on decision-making 
processes and outcomes. We also expect to begin observing changes in individuals’ attitudes and 
beliefs, especially those related to the appropriateness of women’s participation in decision-making 
processes and the legitimacy of TP deliberation fora. It is unlikely that we will observe impact on 
behaviors and conflict at this stage. Endline measurement will occur near the end of the project in 
2023. At endline, if TP or TP+DD have meaningfully affected experienced social conflict, we expect 
all indicators of social conflict, primary measures of attitudes and behaviors, and primary measures of 
behavioral experience of social conflict to decrease, and social cohesion to increase.
The required number of completed cases for the three rounds of data collection are listed below. 
We set the sample size for endline using the power calculations presented in the Appendix. The 
minimum detectable effect sizes (MDES) of our main test comparing individuals from communities 
assigned to TP with individuals from communities assigned to TP+DD is 0.224 standard deviations. 
We calculate midline and baseline target sample sizes assuming a 10 percent attrition rate between 
each round.
(1) Baseline: 3,900 (20 cases per community)
(2) Midline: 3,510 (18 cases per community)
(3) Endline: 3,120 (16 cases per community)
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 17
At baseline, households were randomly selected using a random walk procedure, whereby 
enumerators followed a set of rules for choosing routes to follow in a given settlement and selected 
dwellings to attempt interviews. Within each dwelling, one randomly selected individual 16 and over 
was selected for interviewing. Quotas were used to ensure that the sample resembled the 
population in terms of gender (male and female) and age groups (16-29, 30-49, and 50 plus). Baseline 
respondents will be re-contacted in subsequent rounds.
The questionnaire includes modules related to conflict, social cohesion and trust, integration across 
genders, responses to disasters, development and poverty, migration, views about community and 
regional leadership, women’s leadership, youth’s leadership, other political attitudes and behaviors, 
and personal and household socioeconomic characteristics, such as economic situation, educational 
attainment, and asset ownership.
We plan to supplement the survey data with project implementation data collected by the project 
itself and with data from secondary sources. Project implementation data includes (a) who attends 
the meetings, (b) who among those in attendance participates, and (c) TP staff ratings of community 
engagement. We will use these implementation data to construct process-level outcomes and check 
that the DD intervention produced the anticipated first-order effects. We plan to use data collected 
from reports from local government, police, and other sources, as well as local news organizations 
and government and civil society entities to track social conflict and violence at the community level.
DO NO HARM CONSIDERATIONS
Post-conflict contexts are inherently sensitive, a fact expressly acknowledged in the community￾inclusive design of TP. In the Western Highlands, the known patterns of intrafamily violence, social 
conflict between the young and old, conflict between indigenous groups and extractive industries—
to name only a few—mean that peacebuilding programming and this supporting IE must be 
conducted in a conflict-aware manner. There are four areas of particular concern for both the 
project and IE: (1) communities that will not receive any intervention, (2) municipal structures that 
guide and fund community decision processes, (3) individuals who do not receive program benefits 
within the communities that receive treatment, and (4) traditional and formal authorities that 
function within communities.
The IE team and its partner data collection firm will take several preemptive and ongoing actions to 
mitigate the risk of harm among control communities—those communities that receive no direct 
benefit from the program, but are nonetheless part of the baseline, midline, and endline data 
collection. First, the supervisors, enumerators, and interviewers of our data collection partner will 
receive training in human subject research ethics, including issues of confidentiality and informed 
consent. Second, we will present our data collection activities as an independent study sponsored by 
USAID and will not mention TP or the IE in our field personnel training and data collection activities. 
In doing so, we will minimize the possibility that communities that do not receive a direct benefit 
from TP become aware of the project. Third, our partner will follow their standard protocol of 
securing municipal and community approval before beginning to gather any information in a given 
location. 
A second way that information flows might cause harm could occur at the municipal level. 
Community organizations, when meeting with municipal leaders, may come to find out that TP 
program activities are being undertaken in some communities, but not others. This concern is 
present in nearly all peacebuilding and development programming that cannot enroll every 
community. Both TP and other ongoing programs in the region have well-developed strategies to 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 18
communicate with municipal and community leadership about this concern that build on 
participatory community selection processes. 
Two final concerns exist within communities assigned to be a part of TP program activities. On the 
one hand, program resource constraints necessarily mean that program activities and any associated 
benefits may not accrue evenly across all members of the community. On the other hand, increasing 
the participation of historically marginalized groups risks antagonizing traditional or historical power 
structures within communities. These concerns are inherent to any program activity that TP might 
undertake and not due to the undertaking of this IE, and therefore fall on TP’s implementer to 
address. 
BASELINE DATA COLLECTION
NORC partnered with the Centro de Investigaciones Económicas Nacionales (CIEN) to carry out 
data collection. CIEN is a think-and-action tank headquartered in Guatemala City. CIEN specializes 
in data collection for project evaluation, among other things. All data collection activities were 
conducted under close NORC supervision. 
NORC prepared the survey instrument in English and CIEN translated it into Spanish. The survey 
was administered in Spanish as well as in three Mayan languages: Ixil, Mam, and K’iché. The survey 
was not translated into the indigenous languages at the recommendation of CIEN. Their experience 
indicates that these languages are not sufficiently standardized to ensure full intelligibility between 
dialects. Our strategy was to train fully bilingual enumerators with the Spanish instrument and have 
them work on standard ways to administer the survey in a given indigenous language in language￾specific groups.
The survey instrument went through extensive revisions before the final testing pilot was conducted 
as part of enumerator training. CIEN’s experts answered and subsequently reviewed pen-and-paper 
versions of the survey, resulting in several adjustments prior to the survey being programmed for 
administration using computer-assisted personal interviewing. Specifically, we used the SurveyCTO 
platform programmed onto 6.3-inch Android smartphones. After programming, the survey was again 
reviewed by CIEN and NORC staff to ensure correct skip logic and other programming features, 
and identify any outstanding issues. The programmed survey was then pre-tested in two rural 
communities in the Chimaltenango Department and underwent an additional round of revisions 
before enumerator training.
The primary training exercise for the field personnel was conducted in Guatemala City from 
September 3 to 9, 2019. The first four days were in-class trainings, which involved reviewing each 
questionnaire, interviewing techniques, field protocols, and mock interviews. The next three days 
involved a pilot in the field, a reconciliation and debrief following the pilot, and one day of additional 
training. 
Four teams began data collection on September 10, 2019. Each team consisted of one supervisor 
and four enumerators. CIEN used the following protocol to secure approvals prior to conducting 
fieldwork: In August, an advanced logistics team visited all municipalities to present the study and 
secure approvals to conduct fieldwork. Upon securing approvals at the municipal level and receiving 
contact information for community authorities, the team visited each community to present the 
study and request permission to conduct fieldwork. Most community authorities did not grant 
permission right away, because they had to consult with others in their communities. During 
fieldwork, and prior to the arrival of the enumeration teams in each community, a logistics 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 19
coordinator and supervisors re-contacted community authorities to request permission to conduct 
fieldwork. Permission to collect data was received from 184 of 195 communities. Fieldwork ended 
on October 27, 2019. 
FIELDWORK OUTCOMES
Field teams attempted to conduct interviews in 195 communities, but as mentioned above, they did 
not receive authorization to carry out work in 11 communities. All of these communities were 
located in the Quiché Department (eight in the Chichicastenango Municipality, two in San Juan 
Cotzal Municipality, and one in the San Andrés Sajcabajá Municipality). The total number of contact 
attempts was 6,540. Teams did not find anyone at home in 11.7 percent of cases; the person initially 
contacted in a given household rejected participating in the interview in another 10.9 percent of 
cases; and an additional 11.1 percent of cases did not qualify for the survey because of sample quotas 
or because they did not have household members 16 years old and over. The teams were able to 
identify an eligible survey respondent in 4,336 households and conduct interviews in 3,964 of those 
households. The difference is explained by refusals (192 cases); inability to contact the selected 
respondent (96 cases); disqualifications due to cognitive limitations (16 cases); and other reasons (68 
cases).
Following data collection, an additional 98 records were deleted through data cleaning because of 
completeness and quality issues. The resulting total number of valid cases in the baseline dataset is 
3,866. All data are available in Stata format with a codebook defining variables and their values.
FIELDWORK CHALLENGES
For the most part, fieldwork activities proceeded without major difficulties. The most challenging 
aspect was scheduling, as several authorities changed the dates in which they would allow visits to 
their communities with little advance notice. This required the logistics coordinators and the 
supervisors to contact authorities in other communities to reschedule their visits. It also required 
splitting teams to be able to conduct fieldwork in two communities at the same time and visiting a 
given community on more than one occasion to complete the activities. Teams experienced GPS 
connectivity problems using SurveyCTO, and therefore used handheld GPS devices to capture 
coordinates. 
In addition, as mentioned above, 11 communities did not allow enumeration teams to carry out data 
collection activities. In most cases community authorities were unwilling to participate because 
participation did not entail any material benefits to the community. Another reason authorities gave 
was that their participation in past studies had not yielded any results for them. In line with 
discussions with USAID, NORC and CIEN plan to distribute two-page briefs summarizing survey 
results to all participating communities.3
Lastly, during the course of fieldwork, teams suffered a small number of security incidents. Armed 
robbers stole two tablets and a phone from an enumeration team. And the authorities of two 
communities—Pocohil Primero and Mactzul Sexto in the Chichicastenango Municipality—asked 
 
3 As of the submission of this report, we are preparing the two-page briefs and will share them with USAID for 
feedback. We plan to distribute them as soon as the GOG eases travel restrictions imposed in response to the 
COVID-19 pandemic.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 20
enumerators to immediately leave their communities, threatening them with violence if they failed to 
do so.4
FINDINGS
CONFLICT
To begin to understand conflict, we first ask respondents whether they are aware of conflict having 
occurred within their community in the last 12 months for each of four separate conflict types. 
These types are land; water or water access; natural resources; and governance (i.e., political conflict
or conflict with municipal or community authorities). 
First, we show the distribution of all reported conflict by community in Figure 1. For each 
community, we count the number of individuals who report the incidence of any type of conflict and 
assign the value of the count to the community. The horizontal axis shows the number of individuals 
within a community who report the incidence of any type of conflict. We then plot the number of 
communities with a given number of conflict reports. The figure illustrates that (a) the majority of 
communities had less than five reports of conflict in the last 12 months; (b) many communities had 
between five and ten reports of conflict; and (c) a smaller but substantial number of communities 
had more than ten reports of conflict.
 
4 Enumerator teams visited all communities only with approval from authorities—they granted initial approval 
for the study following visits from the advanced logistics team and confirmed approval during scheduling calls 
and on the day of visits. However, in these two communities, COCODE presidents seem to have provided 
approval unilaterally when their communities’ procedures call for involving other authorities—e.g., auxiliary 
mayors—in the process. These other authorities, together with some community members, expelled the 
teams from the communities. In future data collection efforts we will confirm with authorities that they have 
followed their communities’ usual procedures to grant approval for fieldwork visits.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 21
Figure 1. Reported conflict by community
The following figure repeats this exercise for each conflict type. The behavior of these domain￾specific conflict reports is roughly the same for each domain. In general, the majority of communities 
have few reports of conflict. The long right tail of each distribution shows the few communities that 
have many reports of conflict of a single type. These figures fit well with a model of conflict as a 
relatively rare event that is nonetheless present throughout the region under study.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 22
Figure 2. Reported conflict by community for four conflict types
More nuanced data are recorded within each conflict type. If a respondent reports a given type of 
conflict, we ask them whether the conflict impacted them directly, and whether there was an 
attempt at resolution. For those respondents who report that resolution was attempted, we ask for 
the venue in which this attempted resolution occurred.
Figure 3 illustrates the proportion of respondents who, having reported a given type of conflict, 
subsequently say they were directly impacted by it. We see that as few as 30 percent of the 
respondents say they were directly impacted by governance-related conflict, while slightly over 60 
percent of respondents were directly impacted by water-related conflict. In turn, Figure 4 illustrates 
the proportion of respondents who, having reported a given type of conflict, subsequently say that 
the parties in the conflict attempted to resolve it. Just over 30 percent of the respondents say 
parties attempted resolution for governance-related conflict (i.e., political conflict or conflict with 
municipal or community authorities), while about 64 percent of respondents say parties attempted 
resolution for land-related conflict.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 23
Figure 3. Proportion of respondents directly impacted by conflict, by conflict type
Figure 4. Proportion of conflicts with attempted resolution, by type of conflict
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 24
The following table provides a list of modes/venues where the reported attempts at resolution are 
occurring and their frequencies. By far, the most common modes/venues for attempted resolutions 
are formal institutions at the community and municipal level, along with the courts.
Table 4. Conflict resolution modes/venues
RESOLUTION MODE/VENUE COUNT
Formal community authority (COCODE, others) 346
Judicial system (courts, lawyers) 328
Municipal government (municipal authority/office, COMUDE) 259
Traditional authority (indigenous authorities, others) 89
Informal, person-to-person without outside actor 55
Police 43
National government (including Comisión de Diálogo Presidencial, CPD, 
and Comisión Presidencial de Derechos Humanos, COPREDEH) 
31
Departmental government 27
Church 5
We also have information on the specific impacts of conflicts on our respondents’ behaviors and 
experiences. These questions are open to all respondents and refer to any conflict that their 
community might be experiencing—i.e., they do not refer to a particular type of conflict. Table 5 
shows the number of respondents who say that conflict has impacted them in the way specified by 
each question. The most common conflict impacts are children not attending school and avoiding 
leaving the house alone at night.
Table 5. Conflict impacts
CONFLICT IMPACTS COUNT
Children do not attend school 333
Avoid leaving house alone at night 237
Choose not to participate in social events 168
Avoid using public transportation 151
Choose not to attend religious service 151
Avoid travelling outside community 146
Threat of physical violence 146
Experience hunger 142
Experience damage to property 120
Forced to leave home 62
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 25
Next, we report on questions that we ask to all respondents about several actors who could be 
involved in conflict resolution. We first ask respondents whether or not each of these actors played
a role in resolving conflicts that affected their community in the past 12 months. We label this as 
actor relevance. The boxplots in Figure 5 show the variation in the community-level proportion of 
respondents who believe a given actor is relevant. The lower and upper ends of the whiskers 
respectively indicate the lowest and the maximum proportions excluding any outliers, which are 
represented by dots. The thin lines at the bottom and top of the boxes respectively indicate the first 
and thrid quartiles, while the solid line in the boxes indicates the median. The actors most widely 
viewed as relevant to responding to conflict are community authorities and the COCODE. Actors 
viewed as least relevant are women and youth. The figure shows that there is substantial variability 
in views about actor revelevance across communities.
Figure 5. Respondents who believe actors are relevant for conflict resolution
Of course, labeling an actor as relevant for resolving conflict is not the same as viewing that actor as 
being effective in resolving conflict. Separately, we ask respondents to say whether or not that actor 
was effective in resolving conflict. The boxplots in Figure 6 show the variation in the community￾level proportion of respondents who believe a given actor is effective. The actor most widely viewed 
as effective in resolving conflict is the COCODE. The actor viewed as least effective is the youth, but 
the women as a whole are similarly regarded. The figure also shows that there is substantial 
variability in views about actor effectiveness in resolving conflict across communities.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 26
Figure 6. Respondents who believe actors are effective in resolving conflict
MIGRATION
We ask respondents questions about their own migration history, as well as the migration history of 
the members of their household. Roughly one third of our respondents have not lived in the 
community in which they were interviewed for their entire lives. Of these, more than two thirds 
come from another part of Guatemala, and less than one third come from abroad.5 We gathered 
information on country of origin for these 200 respondents, not reported here. We also ask 
respondents about the reasons that led them to migrate to the community in which they were 
interviewed. We do not present those results here in order to focus on future migration behavior.
Of our respondents, roughly 20 percent say that in the past 12 months they have considered moving 
to another country. Of these, nearly all say their intended destination is Mexico. Fifteen percent say 
they considered moving to another part of Guatemala. Interestingly, the proportions of those 
reporting an intention to migrate corresponds closely to the proportions of those reporting having a 
household member who has migrated in the past 12 months. Roughly 20 percent of respondents 
report that a household member has migrated to another country in the past 12 months. About 11 
percent report that a household member has migrated to another part of Guatemala. As for the 
case of intention to migrate, nearly all respondents told us that the household member who left 
migrated to Mexico.
 
5 We also gather information on department or country of origin for respondents who report having migrated 
as well as on the reasons that led them to migrate to the community in which they were interviewed. We do 
not present those results here in order to focus on future migration behavior.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 27
We ask respondents why they are thinking about migrating. As is typically true in surveys of 
potential migrants, economic motivations emerge as the most widely reported, followed by reasons 
related to family or marriage. Table 6 provides a list of the migration reasons mentioned and their 
frequencies.
Table 6. Migration reasons
DESCRIPTION COUNT
Lack of economic opportunity 664
To look for work 500
Family reason or marriage 107
Community disputes: land 8 
Famine 7 
Property destroyed 4 
Threat of violence 1 
Ethnic conflict 1 
ECONOMIC SITUATION, EDUCATIONAL ATTAINMENT, AND ASSET OWNERSHIP
We ask respondents two standard questions about the economy. We asked first about their views 
of the economy within their community, compared to the previous year. We then asked for the 
same year-over-year comparison, but focused on the respondent’s personal economic situation. The 
results for these questions appear in the following figure. Most respondents say that the economic 
situation in their communities and their own economic situation have not changed.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 28
Figure 7. Perceptions of change in economic situation
Overall levels of education are low. The table shows the proportion of respondents in each 
education category. The categories labeled “High (2)” and “High (3)” refer to the diversificado level.
This is roughly the equivalent of high school, and comes in both a two- and a three-year curriculum. 
Overall, the overwhelming majority of respondents have never had high-school level instruction. We 
collected more finely measured data on years of education, not presented here.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 29
Figure 8. Distribution of educational attainment
As is standard for development surveys, we ask respondents whether or not their household owns 
each of a list of household assets. The frequency of respondents who own each item is presented in 
Table 7. The first point to note is the pervasiveness of mobile technology. Nearly all respondents 
have a cellphone and most respondents also have a television. The least commonly owned items are 
cars, buses, mopeds, and trucks. We plan to use asset ownership information to calculate a proxy 
index for poverty, which is one of TP’s higher-order, longer-term impacts.
Table 7. Household asset ownership
ASSET COUNT
Cellphone 3,554
TV 2,144
Blender 1,700
Cable 1,462
Fridge 1,011
Stove 971
Iron 903
Bicycle 809
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 30
ASSET COUNT
Motorbike 690
Pick-up 427
Car 257
Bus 144
Moped 76
Truck 49
As a quick test of the validity of these asset measures, we report the correlation between education 
and ownership of a refrigerator. Of those with no high school instruction, roughly 18 percent report 
owning a refrigerator. Of those with at least one year of high school instruction or technical 
education, approximately 50 percent report owning a refrigerator.
INTERPERSONAL TRUST AND SOCIAL COHESION
To begin to understand interpersonal trust in TP communities, we simply ask survey respondents to 
state how trustworthy (1) their neighbors; (2) those who live in their community; and (3) those who 
live in other communities are. The pattern of responses paint a remarkable picture of interpersonal 
trust in these communities. By far, the most trustworthy group is those who live in the respondents’ 
communities. Those who live in other communities are the least trustworthy group and 
respondents’ neighbors are somewhere in between. In general, interpersonal trust diminishes with 
the distance between the respondents and the reference group they are asked to assess, so it is 
surprising to find that our respondents trust their neighbors less than those who live in their 
communities. 
There is a divergence in respondents’ patterns of trust for those who live in their community and 
those who live in other communities (i.e., outside of their community). Respondents are much less 
trusting of an unspecified member of another community than they are of individuals who are either 
in their community or are their neighbors. This pattern of beliefs brings into focus the possibility of a 
latent conflict. Respondents have specific, positive feelings about the people who live in their own 
community, but they do not extend these positive feelings toward those who are not part of their 
own community.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 31
Figure 9. Trust in three reference groups
Do individuals’ feelings of trust cohere? That is, are some people very trusting of every other person 
(regardless of whether that other is a community member or a member of another community), 
while other people are very distrustful of everyone? To examine this question, Figure 10 plots each 
respondent’s answers to two questions as a single point. The position along the x-axis indicates the 
individual’s level of trust in those who live in their own communities, and the position along the y￾indicates the level of trust in those who live in other communities. If individuals have some general 
level of trust that they share for every person, without regard to familiarity or co-membership in a 
community, then the preponderance of the data would fall along the plot’s diagonal.
This does not seem to be the case. Instead, the figure shows that respondents can be trusting and 
very trusting of people living in their own communities (those in two rightmost columns) while not 
being trusting of those living in other communities. In addition, we see that respondents who are 
very trusting of those living in other communities (those in the top row) also tend to be very 
trusting of their own community. Respondents who do not trust those living in their communities 
(those in the leftmost column) are not likely to trust those living in other communities.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 32
Figure 10. Trust in those living in own community and other communities
A novel feature of this baseline survey is the use of hypothetical prompts asking community 
members to imagine a negative circumstance and what the community response would be. More 
specifically, we ask community members to imagine an exogenous act of god (a flood, landslide, or 
earthquake) and a threat that affects mostly women in the community (the community is left without 
a midwife) and then ask them three follow-up questions about how the community would respond. 
The questions ask whether the response would be along familial lines or encompassing of the 
community as a whole; along gender lines—separate male or female responses, or all-inclusive 
responses; and along age lines—separate youth or older adult responses, or all-inclusive responses. 
We plan to use these items to construct one measure of social cohesion.
Figure 11 presents the breakdown of responses to the follow-up questions about gender-based 
responses to the act of god and the threat affecting most women. The bulk of respondents say that 
their communities would respond by working together as whole—rather than segregated along 
gender lines—to both types of crises. Interestingly, while there are some respondents who say that 
the response to the act of god would be male-specific, very few say that the response to the 
women-specific threat would be male-specific.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 33
Figure 11. Responses to Two Hypothetical Crises
INTEGRATION ACROSS GENDERS
Motivated by patterns of discrimination against women and gendered segregation in the region, we 
wonder about the social integration of men and women in these communities. To measure this we 
ask two types of questions to all respondents: (1) “Would you say that you generally get along well 
with the (men/women) in the community?”; and (2) “In your spare time, how frequently do you 
spend time with (men/women) in your community who aren’t a part of your family?”
We measure these attitudes because we expect that communities where respondents have positive 
views of and constant interactions with people from both genders are likely to be integrated on 
other social dimensions that we are not able to measure. Figures 12 and 13 respectively present the 
results of questions on how well the respondent gets along with men and women. Figures 14 and 15 
respectively present the results of questions on how frequently the respondent spends time with 
men and women.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 34
Figure 12. Social relationships with men
Figure 13. Social relationships with women
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 35
Figure 14. Social interaction with men
Figure 15. Social interaction with women
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 36
Several patterns are apparent in these responses. The preponderance of respondents say that they 
get along either “well” or “very well” with the men and women in their communities. However, 
even in this reported assessment, there is evidence for gender-based responses. Among those who 
report getting along “very well” with men in the community, male respondents are more than twice 
as likely to report getting along very well than female respondents. When the question turns to ask 
how well people get along with women, the gender-based differences are less acute. Female 
respondents are marginally more likely to note that they get along “very well” with other women 
than male respondents.
When it comes to interaction, the most frequent response is that individuals spend social time with 
others once per week; the most likely explanation is that individuals see one another socially at 
religious services. Patterns also show that individuals spend more time with others who share the 
same gender identity: men more frequently report spending social time with other men daily and 
weekly than do women, and women more frequently report spending social time with other women 
daily and weekly than do men.
VIEWS ABOUT WOMEN’S LEADERSHIP
One question that this project seeks to answer is, “Does increasing the participation of women in 
decision making lead to changes in the norms about who should appropriately participate in 
community governance?” An important precursor to understanding this question is knowing the 
baseline views toward the appropriateness of men’s and women’s participation in leadership. Our 
survey includes several questions to gauge these views.
Figure 16 presents the results of a question asking respondents about their level of agreement with 
the statement “Women in my community participate in making decisions that affect the whole 
community.” This serves as a proxy for gauging the current level of women engagement in 
community affairs. As the figure shows, the vast majority of respondents “somewhat agrees” “or 
“strongly agrees” with this statement, regardless of gender. This first result is a bit surprising given 
prior reports of segregation of women from leadership positions in communities as well as their 
limited participation in community affairs. One possibility is that these reports are overstated. 
Another is that the responses to our survey suffer from social desirability bias, where responses 
reflect what is deemed socially acceptable or desirable, rather than truth. In any case, randomization 
allows us to identify effects even in these circumstances.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 37
Figure 16. Women’s participation in decision making
Next, we present the results to two questions asking respondents about their level of agreement 
with the following statements: “Women are accepted in leadership roles in your community” (Figure 
17) and “There should be more women’s leadership in your community” (Figure 18). In general,
respondents are quite acceptant of women in leadership roles. There are slight differences in 
responses across genders, with females’ views being a bit more positive than those of males. Further, 
respondents generally agree with the statement that there should be more women’s leadership in 
their community. However, there is a more sizable difference between the levels of support for a 
more prominent role for women by gender. Men “strongly agree” with the statement at a 25 
percent lower rate than women, and say they “somewhat agree” with this statement at a rate 10 
percent lower than women.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 38
Figure 17. Acceptance of women’s leadership
Figure 18. Views about more women’s involvement in leadership 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 39
Figure 19 presents the results of a question asking respondents about their level of agreement with 
the statement “When women in my community raise concerns, they are taken seriously.” With this 
question we try to gauge the extent to which women’s voices are taken into account in community 
decision making. As the figure shows, there is a clear consensus that when women’s issues are 
raised, they are taken seriously. However, there is some gap between how well men and women 
believe they are responding to these concerns. A greater proportion of men than women “strongly 
agree” that women’s concerns are taken seriously, while a greater proportion of women “somewhat 
agree.” This difference is slight, but signals either some difference in the expectations for what it 
means to take a concern seriously, or what an appropriate response to a serious concern is or 
should be.
Figure 19. Views about women’s concerns being taken seriously
Finally, we report responses to a question asking respondents about their level of agreement with 
the statement “Women in my community are as capable of being leaders as men.” As Figure 20 
shows, views are once again quite positive; there is a clear consensus that women are as capable as 
men when it comes to leadership. There are some differences in responses across genders, with 
females’ views being a bit more positive than those of males. Differences, however, are not as stark 
as one would expect in a context where women have traditionally been excluded from leadership 
positions. 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 40
Figure 20. Views about women’s leadership ability
We also ask respondents to share their views on the barriers to women’s participation in 
community governance. We present respondents with a number of statements summarizing 
potential reasons and ask them to state their level of agreement. The potential reasons we present 
them are: (1) lack of political skills and training among women; (2) cultural norms and stereotypes 
about the role of women; (3) cultural norms that men are better suited to being leaders than 
women; (4) lack of safety and respect for women in the political sphere; (5) formal laws and policies 
preventing the participation of women; (6) lack of solidarity among women; (7) lack of education and 
literacy among women; and (8) women’s greater responsibilities at home.
We do not present figures with the results to economize space. In general, the majority of 
respondents “somewhat agree” or “strongly agree” with all the reasons provided—in other words, 
respondents would not discount any of the reasons provided. The reasons that most respondents 
“strongly agree” with are: lack of safety and respect for women in the political sphere; formal laws 
and policies preventing the participation of women; lack of education and literacy among women; 
and women’s greater responsibilities at home. Importantly, difference across genders in responses is
only modest.
CIVIC ENGAGEMENT
Our survey also includes a number of questions gauging civic and political engagement. These 
questions ask respondents to indicate the frequency at which they participate in meetings of various 
types of social and political groupings or organizations, including religious services, community 
committees, political parties, and so forth. Instead of presenting those responses, here we focus on 
two items that we ask immediately following the meeting questions. First we ask respondents 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 41
whether they would have wanted to attend meetings but were unable to. If they say they were 
unable to attend, we ask them the reason.
As is shown in Figure 21, the preponderance of respondents do not believe that there are things 
that keep them from participating in meetings if they wanted to. However, approximately one-third 
of those surveyed say that there are barriers to their participation in meetings. One important point 
about the way this question is asked bears description: This question asks, “Are there times that you 
have wanted to participate in meetings, but were not able to?” There are two ways that an individual 
could respond, “No.” First, they might never have wanted to participate in meetings; second, they 
might have wanted to participate and have always been able to participate. In particular, this would 
be challenging for the cases where one group has no expectation of participating.
Figure 21. Inability to attend social and political meetings 
Among those who are not able to participate in meetings, far and away the most common reason 
identified is a lack of time. The way that this question was asked, it is possible for a respondent to 
reply to more than a single category. As a result, individuals could both respond, “I don’t have time,” 
and, “I had to work at home,” and they could have in mind the very same obstacle that kept them 
from participating.
In Figure 22, we remove responses related to “‘no time” to attend so that it is possible to focus on 
the more specific reasons. We also present responses disaggregated by gender. Considerably more 
men than women report that work at home kept them from participating, a surprising result on its 
face. Other results are consistent with a cultural sense of appropriateness for men’s and women’s 
participation in these events: more women than men identified that it would be inappropriate for 
them to participate and that it might be dangerous for them to participate. 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 42
Figure 22. Reasons for not attending meetings
CONCLUSION
This study lays the groundwork to assess the overall effectiveness of TP activities at the community 
level, as well as whether an intentional focus on the participation of women in community 
deliberations can further contribute to reducing conflict and increasing cohesion. We seek to expand 
the evidence base by demonstrating that women’s involvement in local decision-making processes 
can not only contribute to development outcomes but also to peacebuilding activities. In this 
baseline report, we describe the rationale and background for the study and the DD intervention, 
asesoras de cambio, our experimental approach to measure impact, the baseline data collection 
process, and our analysis of the baseline data.
We draw several conclusions from our baseline analysis:
• There is considerable heterogeneity in the distribution of conflict within the Western Highlands 
region. In some communities there is little evidence of conflict presence, while in others there is
evidence of acute levels of conflict. 
• Conflict related to land and conflict related to water were more likely to be present in 
communities than conflict related to governance and conflict related to natural resources.
• Most survey respondents state that COCODES and community leadership play a leading role in 
conflict resolution. In stark contrast, very few respondents state that women and youth play a 
role in the conflict resolution. Few respondents feel that these groups (or any other group) is 
actually effective in resolving conflict.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 43
• Most community members believe that members of their community are somewhat or very
trustworthy. This indicates a general willingness to trust others. However, there is lower trust 
for neighbors. And, there is a notable lack of trust between members of distinct communities: 
respondents most frequently said that members of other communities are not trustworthy.
• Trust in respondents’ own communities and other communities is relatively tightly aligned. 
When respondents distrust those in their own community, they are also very unlikely to trust 
those who live in other communities. When respondents trust those in their own community, 
there is roughly a 50-50 chance that they will also trust other communities. 
• There is a broadly held view among both men and women that women have a role to play in 
leadership in the community. However, other evidence indicates that women’s actual 
involvement in leadership and civic organizations is more limited than men’s. 
As of the writing of this report, the IE has been cancelled. The IE team does not know whether the 
asesoras de cambio intervention will be implemented as designed by the TP team or supplemented 
with recommendations the IE team offered during TP’s 2019/2020 work plan review process. 
Further, the IE team does not know whether the asesoras de cambio intervention will be 
implemented in line with the IE design—in only 65 communities randomly assigned to the TP+DD 
treatment arm. We understand that current funding restrictions prevent USAID/Guatemala from 
contributing funding to IE activities in the short term.
Our hope is that additional funding for the IE becomes available toward the end of the project in 
2023, in time to carry out endline data collection, analysis, and reporting activities. We remain 
hopeful for two reasons. First, TP’s monitoring, evaluation, and learning plan includes performance 
indicators whose quantification requires survey data collection in TP communities. Assuming that the 
project is planning to use funds to this end and a minimal approach to data collection consisting only 
of the survey, we would only need additional funding to collect data in the 65 control communities 
plus data analysis and reporting activities. Second, full U.S. foreign assistance is likely to resume in the 
medium term, as governments in the Northern Triangle undertake action to address illegal 
immigration. 
Should the implementer abide by the full randomization schedule and additional funding become 
available, we would be able to estimate the impact of the asesoras de cambio intervention by 
comparing communities assigned to TP+DD with communities assigned to TP. We would also be 
able to estimate the impact of the standard TP project by comparing communities assigned to TP 
with communities assigned to the control group. To assess DD’s expected effects on decision￾making processes and outcomes at midline, we propose analyzing project monitoring data, such as 
event attendance records, as well as coding and analyzing CRA and CV documents.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 44
REFERENCES 
Adams, Tani. 2015. “Legacies of Exclusion: Social Conflict and Violence in Communities and Homes 
in Guatemala’s Western Highlands.” Guatemala Conflict Vulnerability Assessment. Final Report. 
Prepared by Democracy International, Inc. for USAID.
Alesina, Alberto, Reza Baqir, and William Easterly. 1999. “Public Goods and Ethnic Divisions.” The 
Quarterly Journal of Economics 114(4): 1,243-1,284.
Allport, Gordon W. 1958. The Nature of Prejudice: A Comprehensive and Penetrating Study of the Origin 
and Nature of Prejudice. Addison-Wesley.
Bhavnani, Rikhil R. 2009. “Do Electoral Quotas Work after They Are Withdrawn? Evidence from a 
Natural Experiment in India.” American Political Science Review 103(1): 23-35.
Bornman, E., and J. C. Mynhardt. 1991. “Social Identification and Intergroup Contact in South Africa 
with Specific Reference to the Work Situation.” Genetic, Social, and General Psychology Monographs 
117: 437–62. 
Broockman, David, and Joshua Kalla. 2016. “Durably Reducing Transphobia: A Field Experiment on Door￾to-Door Canvassing.” Science 3352(6282): 81–87. 
Chattopadhyay, Raghabendra, and Esther Duflo. 2004. “Women as Policy Makers: Evidence from a 
Randomized Policy Experiment in India.” Econometrica 72(5): 1,409-43.
Eskilson, A. 1995. Trends in Homophobia and Gender Attitudes: 1987–1993. Presented at the Annual 
Meeting of American Sociological Association, 90th, Washington, DC. 
Fearon, James, Macartan Humphreys, and Jeremy Weinstein. 2015. “How Does Development 
Assistance Affect Collective Action Capacity? Results from a Field Experiment in Post-Conflict 
Liberia.” American Political Science Review 109(3): 450-69.
Franceschet, Susan, and Jennifer M. Piscopo. 2008. “Gender Quotas and Women's Substantive 
Representation: Lessons from Argentina.” Politics and Gender 4(3): 393-425.
Gerber, Alan and Don Green. 2015. Field Experiments: Design Analysis and Interpretation. WW 
Norton. 
Herek, G. M., and J. P. Capitanio. 1996. “Some of My Best Friends:” Intergroup Contact, Concealable 
Stigma, and Heterosexuals’ Attitudes toward Gay Men and Lesbians.” Personality and Social Psychology 
Bulletin 22: 412–24. 
Mansuri, Ghazala, and Rao Vijaynedra. 2012. Localizing Development: Does Participation Work? 
Technical Report World Bank, Washington, D.C.
McKay, S., and J. Pitman. 1993. “Determinants of Anglo-Australian Stereotypes of the Vietnamese in 
Australia.” Australian Journal of Psychology 45: 17–23. 
Miller, Grant. 2008. “Women's Suffrage, Political Responsiveness, and Child Survival in American 
History.” The Quarterly Journal of Economics 123(3): 1,287-1,327. 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 45
Olken, Benjamin. 2010. “Direct Democracy and Local Public Goods: Evidence from a Field 
Experiment in Indonesia.” The American Political Science Review 104(2): 243-267. 
Pande, Rohini. 2003. “Can Mandated Political Representation Increase Policy Influence for 
Disadvantaged Minorities? Theory and Evidence from India.” The American Economic Review 93(4): 
1,132-1,151.
Pettigrew, T. F. 2006. "A Meta-Analytic Test of Intergroup Contact Theory." Journal of Personality and 
Social Psychology 90(5): 751–83.
Piscopo, Jennifer M. 2014. “Female Leadership and Sexual Health Policy in Argentina.” Latin American 
Research Review 49(1): 393-425.
Riordan, C. 1987. “Intergroup Contact in Small Cities.” International Journal of Intercultural Relations 
11: 143–54.
United Nations Secretary-General (2018). Report of the Secretary-General on Women and Peace 
and Security. United Nations, New York.
Waldron, Jeremy. 1999. Participation: The Right of Rights. Oxford University Press.
Wagner, U., M. Hewstone, and U. Machleit. 1989. “Contact and Prejudice Between Germans and 
Turks.” Human Relations 42: 561–74.
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 46
APPENDIX: POWER CALCULATIONS
We propose to undertake a minimum of 16 surveys in each of 195 communities (130 to be split 
evenly between treatment arms and 65 to serve as control) to evaluate the impact of TP and 
TP+DD. The resulting total number of surveys is 3,120. Two core comparisons are of interest 
within this design. On the one hand, we measure individual attitudes and beliefs in communities 
assigned to treatment and control arms. On the other hand, we measure community level 
outcomes, e.g., occurrences of social conflict that rise to the level of physical violence. All 
measurements are described in Table 1 and Table 2. Here we describe the statistical power 
associated with these comparisons.
Often, in laboratory experiments, minimum detectable effect sizes (MDES) that are less than 0.2 are 
characterized as “small” or desirable, meaning that the experiment is well powered to detect even 
small effects. MDES between 0.2 and 0.6 are characterized as “medium” effect sizes, meaning that an 
experiment is well powered to detect only relatively larger effects.
Comparison 1: What are the differences in individual-level outcomes (e.g., attitudes and beliefs) when 
comparing individuals who live in communities assigned to TP with individuals who live in 
communities assigned to TP+DD? Alternatively, what are the differences when comparing 
individuals who live in communities assigned to TP with individuals who live in communities assigned 
to control? 
In Table A1 we report the power analysis for the test of differences in outcomes between individuals 
who live in communities assigned to TP and individuals who live in communities assigned to TP+DD. 
Two features make measuring an effect more difficult. First, there are necessarily fewer communities 
to power the test, as it is limited to the 130 TP or TP+DD communities. Second, because all 
communities are receiving TP, there is less contrast between the comparison groups: differences in 
effect at this level will be caused solely by the diverse decision makers interventions. Despite these 
limitations, the MDES for this comparison is 0.224 under plausible assumptions, which is just short of 
being a “small” effect.
Table A1. Statistical Power for Individual-level CBPT vs. CBPT+DD Tests
ASSUMPTIONS COMMENTS 
Alpha Level (α) 0.05 Probability of a Type I error
Two-tailed or One￾tailed Test? 2 
Power (1-β) 0.80 Statistical power (1-probability of a Type II error)
Rho (ICC) 0.20 Proportion of variance in outcome that is between clusters
P 0.50 Proportion of Level 2 units randomized to treatment:
JT / (JT + JC) 
R1
2 0.10 Proportion of variance in Level 1 outcomes explained by Level 1 
covariates
R2
2 0.20 Proportion of variance in Level 2 outcomes explained by Level 2 
covariates 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 47
ASSUMPTIONS COMMENTS 
g* 4 Number of Level 2 covariates
N per cluster 16 Mean number of Level 1 units per Level 2 unit (harmonic mean 
recommended)
Number of Clusters 130
M (Multiplier) 2.82 Computed from T1 and T2
T1 (Precision) 1.98 Determined from alpha level, given two-tailed or one-tailed test
T2 (Power) 0.84 Determined from given power level
MDES 0.224 Minimum Detectable Effect Size
Comparison 2: What are the differences in individual-level outcomes (e.g., attitudes and beliefs) when 
comparing individuals who live in communities assigned to any form of TP (i.e., TP and TP+DD) 
with individuals who live in communities assigned to control? 
As we report in Table A2, to be able to detect effects at the household or individual level between 
communities that receive any TP programming (core TP or TP+DD) and control communities, the 
treatment must cause a standardized difference in outcomes of at least 0.193. The increase in power 
achieved by a comparison against control communities is achieved through a limitation of the 
questions that can be asked, as it provides information regarding the overall effectiveness of 
receiving any TP programming.
Table A2. Statistical Power for Individual-level Any CBPT vs. Control Tests
ASSUMPTIONS COMMENTS 
Alpha Level (α) 0.05 Probability of a Type I error
Two-tailed or One￾tailed Test? 2 
Power (1-β) 0.80 Statistical power (1-probability of a Type II error)
Rho (ICC) 0.20 Proportion of variance in outcome that is between clusters
P 0.66 Proportion of Level 2 units randomized to treatment:
JT / (JT + JC) 
R1
2 0.10 Proportion of variance in Level 1 outcomes explained by Level 1 
covariates
R2
2 0.20 Proportion of variance in Level 2 outcomes explained by Level 2 
covariates 
g* 4 Number of Level 2 covariates
N per cluster 16 Units observed in each cluster 
Number of Clusters 195
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 48
ASSUMPTIONS COMMENTS 
M (Multiplier) 2.82 Computed from T1 and T2
T1 (Precision) 1.97 Determined from alpha level, given two-tailed or one-tailed test
T2 (Power) 0.84 Determined from given power level
MDES 0.193 Minimum Detectable Effect Size
Comparison 3: What are the differences in community-level outcomes (e.g., rates of physical violence) 
when comparing communities assigned to TP with communities assigned to TP+DD? Alternatively, 
what are the differences when comparing communities assigned to TP with communities assigned to 
control? 
Beyond attitudes and beliefs, these outcomes are indicators about the functioning of local 
government, peacebuilding, and conflict resolution instruments. Importantly, they are also 
community-level indicators of acutely experienced conflict, for example, a comparison of the 
rate of violent conflict in treatment communities compared to control communities. Measuring 
community-level outcomes is a difficult task with only 65 communities in each treatment 
condition. As we report in Table A3, the DD interventions must cause a standardized difference 
in outcomes of at least 0.393 to be detected under this design.
Table A3. Statistical Power for Community-Level CBPT vs. CBPT+DD Tests
ASSUMPTIONS COMMENTS
Alpha Level (α) 0.05 Probability of a Type I error
Two-tailed or One-tailed 
Test? 2 
Power (1-β) 0.80 Statistical power (1-probability of a Type II error)
P 0.50 Proportion of the sample randomized to treatment: nT / (nT + nC) 
R2 0.20 Percent of variance in outcome explained by covariates
k* 0 Number of covariates used
n (Total Sample Size) 130 
M (Multiplier) 2.50 Computed from T1 and T2
T1 (Precision) 1.65 Determined from alpha level, given two-tailed or one-tailed test
T2 (Power) 0.84 Determined from given power level
MDES 0.393 Minimum Detectable Effect Size
Comparison 4: Finally, what are the differences in community-level outcomes (e.g., rates of physical 
violence) when comparing communities that are assigned to any form of TP (i.e., TP and TP+DD) 
with communities assigned to control? 
CONTRACT NO. GS-10F-0033M / ORDER NO. AID-OAA-M-13-00013 / DRG-LER I TASKING N066
USAID.GOV USAID/GUATEMALA PEACEBUILDING PROJECT IE: BASELINE REPORT | 49
This comparison is slightly better powered than the preceding one and can similarly be used to 
address questions about the reductions in conflict and increase in social cohesion that arises as a 
result of any form of peacebuilding programming. While this is a complex comparison, because the 
treatment arms contain distinct forms of intervention, the comparison does provide information 
about the overall effectiveness of the program activities. As we report in Table 6, to detect 
community-level effects between treatment communities and control communities, the treatment 
must cause a standardized difference in outcomes of at least 0.338. 
Table A4. Statistical Power for Community-level Any CBPT vs. Control Tests 
ASSUMPTIONS COMMENTS
Alpha Level (α) 0.05 Probability of a Type I error
Two-tailed or One-tailed 
Test? 2 
Power (1-β) 0.80 Statistical power (1-probability of a Type II error)
P 0.66 Proportion of the sample randomized to treatment: nT / (nT + nC) 
R2 0.20 Percent of variance in outcome explained by covariates
k* 0 Number of covariates used
n (Total Sample Size) 195 
M (Multiplier) 2.50 Computed from T1 and T2
T1 (Precision) 1.65 Determined from alpha level, given two-tailed or one-tailed test
T2 (Power) 0.84 Determined from given power level
MDES 0.338 Minimum Detectable Effect Size
U.S. AGENCY FOR INTERNATIONAL DEVELOPMENT 
1300 Pennsylvania Avenue, NW
Washington, DC 20523